[RADIOCARBON, VOL. 4, 1962, P. 13-261 GEOLOGICAL SURVEY OF CANADA RADIOCARBON DATES I* W. DYCK and J. G. FYLES Geological Survey of Canada, Ottawa, Canada INTRODUCTION This paper reports the first ages determined in the C14 Dating Laboratory of the Geological Survey of Canada. The C14 dating program of the Geological Survey is a cooperative project; geologists of the Pleistocene Section assess and select samples for dating, and the Isotope and Nuclear Research Section, under Dr. R. K. Wanless, developed and operates the laboratory and calculates ages. The first part of this paper, devoted to sample preparation, counting procedure, and interlahoratory check dates was prepared by the first author, who built and operates the laboratory. The date list was compiled by the second author from descriptions of samples and interpretations of dates provided by various col- lectors. Most samples analyzed so far have originated within the Geological Survey. Preparations for the development of a C14 laboratory within the Depart- ment of Mines and Technical Surveys, Ottawa, were begun by Dr. C. Lapointe, Nuclear Radiation Section, Mines Branch. At an early stage the project was turned over to the Geological Survey and active construction of the present laboratory was begun during 1959. By early 1961 one proportional-counting tube with highly satisfactory counting characteristics was in operation and systematic dating commenced. The age determinations reported here were completed from January to November 1961. ACKNOWLEDGMENTS Acknowledgement is made to Dr. W. J. Broecker of the Lamont Geological Observatory for suggestions regarding some facets of the laboratory develop- ment, and to Dr. K. J. McCallum, Department of Chemistry, University of Saskatchewan, for providing the samples used for the check dates GSC-20 and 21. The proportional-counting tubes were fabricated by A. G. Meilleur, Su- perintendent of the Geological Survey Research and Development Shop. APPARATUS AND PROCEDURE Sample Preparation All organic samples, unless noted otherwise, were treated with hot 4% NaOH, hot 2N HCI, and washed with hot water. Samples were burned in a stream of oxygen and the released CO2 purified by passage through the following chemicals and traps: hot CuO, dil. H2SO4, 0.1 N AgNO3, H25O4-CrO3 sol'II, drierite, an acetone-dry ice trap, hot Pt- asbestos, and hot Ag-wool. Radon was separated from CO2 by fractional distil- lation as described by de Vries (de Vries, 1957). Final purification of CO2 was achieved by passage through Mg(C104) 2 and hot, freshly regenerated copper. * Published by permission of the Director, Geological Survey of Canada, Ottawa, Canada. 13 14 W. Dyck and J. G. Fyles Shells were cleaned with a stiff brush and water before and after removal of 10% of outer layers with HCI. CO2 was liberated with H3P04. Counting Apparatus The counting apparatus is situated in the sub-basement of the eight story Geological Survey building. Samples were counted in a 2 L copper, proportional counter. The end plates and quartz insulators were glued in place with araldite. This counter is operated within a shield comprised of 8 in. of cast iron, 4 in. of paraffin, 21 G.M. tubes and a stainless steel vessel providing a 1 in. layer of Hg. When filled with CO2 to a pressure of 150 cm of mercury a background count of 1.5 counts,/min and a net modern-wood count of 20.2 counts/min were recorded. Experiments with several counters suggest the low background was ob- tained by machining the inside surface of the counter tube. Counting Procedure With a few exceptions the counter was operated at two atmospheres. Dur- ing the first eight months of operation one background and one modern wood count were made each week and each unknown sample was counted at least twice for 20 hrs. Recently a less rigorous counting procedure was introduced as follows : background, two unknowns, modern wood, two unknowns, back- ground. Unknowns counted over a weekend, unless very old, were not counted twice. A new background and standard gas were prepared every six weeks. Each counter filling was checked with an external pitchblende source before and after the counting period in order to detect possible shifts in the operating voltage. Variations of up to 15 v were observed between different counter fill- ings, suggesting small differences in the concentration of electronegative im- purities in the sample gases. A similar shift was observed after a gas was used four or five times. Calculations and Errors Ayll ages reported were calculated using the corrected activity of 105 yr old Douglas Fir wood from Vancouver Island as the reference standard. Two gas preparations of NBS oxalic-acid standard compared to two of the Douglas Fir wood gas preparations gave the following relationship: 0.945 A' = A where A' is the counting rate of the oxalic-acid gas and A the counting rate of the wood corrected to zero age. The value of 5568 ± 30 yr for the half life of C14 was used in these cal- culations. Upper age limits were calculated from values obtained by adding four standard deviations to statistically insignificant counting rates. No corrections for possible isotopic fractionation were applied. Although the C13/C12 ratio for shells and terrestrial plants is different, ocean mixing rates are apparently such that the corresponding difference in the C14/C12 ratio is not observed for shallow-water sea shells (Craig, 1954). Hence the Geological Survey of Canada Radiocarbon Dates 1 15 same standard was used for all samples. Close agreement between GSC-24 and GCS-38 lends support to this argument. Errors reported were based on statistical variations observed. Counts outside the probability range, as predicted from the randomness of radioactive decay, were discarded; i.e. if one of three counts was more than two standard deviations away from the average. During the past 11 months six such anomal- ous counts were recorded. Results of Check Samples Determinations of ages of check samples listed in Table 1 were carried out over a period of eight months. Satisfactory agreement between the ages determined in this laboratory and those from other laboratories is evident. However, slight trends appear to be present; three of the four Isotopes Inc. ages are older than the GSC ages, while all three Lamont ages are slightly younger. TABLE 1 Comparison of ages of identical sample fractions dated by various laboratories GSC Laboratory Other Laboratories Sample Sample Reference No. Age (yr) No. (yr) GSC-1 12,400 ± 200 L-391 D 250 V 12,000 ± 450 Isotopes I GSC-2 7600 ± 150 5-99 120 III peat GSC-4 10,190±120 I(GSC)-185 II GSC-5 10,140 ± 160 I ( GSC) -2 350 I GSC-13 29,500±800 L-424C V GSC-14 26,000±600 L221A V GSC-20 4275 ± 100 S-25 210 II GSC-21 10,400±140 S-81 GSC-37 12,600 ± 170 I (GSC) -248 175 II shells The peat sample GSC-4 was dated three times and a summary of the sample treatment and results is shown in Table 2. TABLE 2 Summary of treatment and results of sample GSC-4 Sample No. of preparation counts treatment 1 2 40% of sample treated with NaOH Total sample treated with HCl 1 Total sample treated with base and 140 acid 3 2 No chemical pretreatment 150 Although the age appears to decrease with increasing chemical pre- treatment, the variation is not significant in view of the errors of the ages. The first count of the second preparation gave an age of 9860 ± 160 yr but was 16 W. Dyck and J. G. Fyles omitted because of the presence of some radon in the sample. Also the first count of the third preparation was omitted. Although no radon was detected in this preparation the counting rate was more than two standard deviations above average. Sample GSC-5 was counted once at a reduced pressure of 147 cm and once mixed with 24.4% coal gas. The counts gave the ages 10,200±190 yr and 10,090 ± 130 yr respectively. Sample GSC-21 and its unpublished Saskatchewan date were provided by Dr. K. J. McCallum. As indicated in a letter from Dr. McCallum (November 28, 1961), this sample is from the Scrimbit site, Saskatchewan (49° 46' N Lat, 105° 11' W Long). The wood is from a depth of 15 ft above till, and was taken from a large root of a stump nearly 4 ft high, which appeared to be pre- served in its original upright position. Dating of a series of samples from this site is being undertaken in the Saskatchewan Laboratory. SAMPLE DESCRIPTIONS (GEOLOGICAL SAMPLES) SOUTHERN CANADA Arranged East to West GSC-11. Clarenville, Newfoundland 3610 ± 100 Peat from the bottom of a bog deposit at the head of SW arm of Random Sound, 10 mi S of Clarenville, Newfoundland (48° 02' N Lat, 53° 48' W Long). Sample collected with Hiller peat sampler 10 ft below bog surface at the contact of peat with underlying outwash gravel at alt 46 ft. The outwash is graded to a former sealevel at alt 35 to 40 ft. Coll. 1960 by E. P. Henderson, Geol. Survey of Canada, Ottawa. Comment: sample was dated to obtain a minimum age for the outwash; it is considerably younger than anticipated. In view of the much greater age of bog-bottom samples from the Avalon Peninsula to the SW (L-391I, 7400 ± 150, Lamont V; I (GSC) -4, 8420 ± 300, Isotopes I), the dated peat must be considerably younger than the underlying outwash. NaOH-leach was omitted from pretreatment. GSC-18. Siegas, New Brunswick 9820 ± 130 Wood from an organic layer, containing moss and fresh-water mollusc shells, lying within gravel 18 ft below ground on W bank of Siegas River 0.3 mi from its junction with St.